Lions and Tigers and Bit Depth….. Oh My

This will be just a quickie today. There’s a lot of confusion in digital photography circles about the terms Dynamic Range and Bit Depth and how they’re related. Some think that they’re completely interconnected. They’re not. It gets even more confusing when HDR is brought into the mix. Let’s try to clarify.

First a bit of photographic geek-speak.  Dynamic range is the total range of brightness that a photographic image can display.  It’s typically expressed by the number of ‘stops’ of light.  Bit depth is the number of colours or shades that can be found in between the two extreme ends.  The two images below show this.  Both have the same dynamic range.  The second one has a higher bit depth (more levels in between).

Dynamic Range/Bit Depth (click for larger version)

Dynamic Range/Bit Depth (click for larger version)

OK, so that’s the geeky part of the exercise.  Let’s put it into terms that are more easily understood.

I read an analogy recently that I think sums it up well.  Think of a building.  That building has 8 floors.  In between each floor are stairs.  The number of floors in the building is the dynamic range.  The number of stairs in between the floors is the bit depth.  There might be 8 stairs, there might be 16 stairs.  But there are still only 8 floors. Each bit represents a shade or colour in between the two ends.  I’ve used black/white/grey here but the same thing applies to colour as well.  With colour we have 3 channels or colours (Red, Green & Blue) and the various combinations of colours can combine to give us the many millions or billions of colours that are often talked about.

Now let’s take the analogy a bit further to describe HDR.  First, a bit more photographic geek-speak.  HDR or High Dynamic Range images contain 32 bits (normal images are 8 or 16, RAW images can be 12, 14 or 16).  That’s a lot of extra bits in HDR – and a lot wider range of colours/tones.  The other difference with HDR is that these bits are what are called ‘floating point’.  What’s that mean?  It means there can be decimals rather than just discrete numbers. True HDR images have wider dynamic range than single images as well.

Back to our building analogy.  An HDR building might have 17 floors (we can’t use all those floors so we tonemap or edit the image to compress everything into a ‘normal’ 8 or 10 floor building).  Rather than discrete stairs between each floor, the HDR building has escalators that allow for smoothly variable  movement between floors.  Like the example image below.

HDR Dynamic Range/Bit Depth (click for larger version)

The big difference with HDR is that the 32 bit floating point numbers used to represent colour (as opposed to integer numbers with LDR images) can accommodate negative numbers and very, very high numbers.  As a result there’s essentially no limitation on the depth of black that can be rendered or the brightness of white.  This is why HDR formats can handle a much wider dynamic range than LDR formats.

So that’s my quickie explanation of dynamic range and bit depth.  Does it work for you?  If you were confused before does it make things clearer?  Let me know!

Addendum:  Bit depth, as we can see from the examples above has an impact on the smoothness of transition between tones.  More steps means a smoother transition.  This also carries through to editing.  Editing in 16 bit vs. 8 bit (or 32 bit vs. 16 bit) makes a difference as well.  The step wedges above were created in 16 bit mode then dropped to 8 bit for display here.  Note that the lines between the grey steps are smooth.  The image below is a gradient step wedge that was created in 8 bit.  The method used to create this and the ones above was to create the gradient then use the Image>Adjustments>Posterize command to create the steps.  Notice in the 8 bit version below how unsmooth the lines are between the steps.  If you click to open the full size image, you’ll also see that there is colour noise along those separations as well.  This is a very simple way to illustrate the benefit of higher bit depth editing and the retention of quality that results.

8 bit Gradient (click for larger version)



Powered by Facebook Comments